Although standard modalities of treatment for melanoma and colorectal cancer (i.e. radiation, chemotherapy, and surgery) have had some impact on the course of these disease, it is clear from the substantial death rate from progressive tumor growth that new, improved approaches are needed. Immunotherapy is a strategy that has gained much interest as a possible fourth modality for the treatment of cancer. The genetic modification of human tumors for the expression of immunostimulatory gene products (namely cytokines) holds promise as a new approach for active immunotherapy of cancer and for the isolation of immune effector cell populations for use in adoptive immunotherapy protocols. To date, the focus of much attention has been directed to the interferons, interleukins, and hematopoietic colony stimulating factors. However, based on substantial published data in animal tumor models, it is unlikely that any of these particular cytokines secreted by the injected gene-modified tumor cells will, by itself, have a profound effect in mediating the complete regression (or ultimately resulting in the cure) of advanced, poorly-immunogenic tumors in patients. We propose to enhance the activity of cancer vaccines through a combined strategy that employs the potent antigen- presenting capacity of dendritic cells coupled with the ability of chemokines to mediate potent chemotaxis of immune effector cells. Our rationale for this new therapeutic approach is based on experimental data and methodologies that we have developed in both murine and human tumor models. To accomplish our goal, we propose the following three (3) Specific Aims: 1. To evaluate in vitro the capacity of human dendritic cells (DC) to detect t cell specific responses to autologous melanoma and colorectal carcinoma; 2. To determine in vitro the capacity of chemokine-secreting cells combined with DC pulsed with autologous tumor to detect, attract, and augment specific, antigen-reactive T cells; 3. To initiate a phase I clinical trials in advanced cancer patients that employ vaccinations with DC pulsed with autologous tumor. The overall goal of our research effort will be to develop a new, innovative molecular vaccine strategy for eventual use in cancer patients that employs autologous chemokine gene- modified fibroblasts combined with tumor-pulsed dendritic cells to both recruit/concentrate relevant immune populations at the vaccination site as well as to activate the recruited T cells by potent presentation of tumor-associated antigens.